Currently the download images code uses a timestamp to create the JPG filename: https://github.com/KewBridge/specimens2illustrations/blob/a72901971abc3e17c742ebe59dad48dffa1cd8ca/xml2illustrationdata.py#L30
It would be better if these were stored under a predictable name generated from the fig_url value. This would mean that subsequent script runs would overwrite existing files rather than accumulating many extra ones and work better with dependency management when we include segmentation (#11 ) in the makefile process.
Currently the download images code uses a timestamp to create the JPG filename: https://github.com/KewBridge/specimens2illustrations/blob/a72901971abc3e17c742ebe59dad48dffa1cd8ca/xml2illustrationdata.py#L30 It would be better if these were stored under a predictable name generated from the fig_url value. This would mean that subsequent script runs would overwrite existing files rather than accumulating many extra ones and work better with dependency management when we include segmentation (#11 ) in the makefile process.